What is it about?

How do people respond when a conversational AI tries to change their mind? We examined whether people respond differently to persuasive attempts from what they think is an AI versus another person. Participants engaged in real conversations about an opinion they held on topics related to lifestyle or consumption choices. They were led to believe that they were interacting with either an AI or another human participant, although in both cases the responses were actually generated by the same AI model. We programmed this AI model to try to persuade the participant towards a different opinion on the topic. When participants thought they were talking to another person, they were more likely to argue against the opposing viewpoint. However, when they thought they were talking to an AI, they argued less, and ultimately became more persuaded by the AI’s position. We found evidence that this happened because people tend to see AI as less persuadable and thus argue less with it. Because they become less argumentative, they do less to defend or bolster their own initial position. These results highlight a unique persuasive advantage for conversational AI: an opposing viewpoint may encounter less resistance when it is thought to come from an AI versus another person.

Featured Image

Why is it important?

AI models are playing a growing role in influencing our preferences and choices. We consult with AIs on matters ranging from how to patch up a strained friendship to which socks to buy. Moreover, many AI developers are beginning to introduce advertising into their AI platforms. Our findings suggest that when AIs take on this role of persuader, they may be more effective than their human counterparts: people argue less with them and, by extension, can be more persuaded by them. As such, our research shows that these models we increasingly turn to for advice can have an advantage when changing our minds. Whether that’s a good or bad thing depends on your perspective (until an AI changes it).

Read the Original

This page is a summary of: From people to prompts: Understanding susceptibility to persuasion in conversations with artificial intelligence., Journal of Personality and Social Psychology, September 2026, American Psychological Association (APA),
DOI: 10.1037/pspi0000531.
You can read the full text:

Read

Contributors

The following have contributed to this page